Thiazolidinedione Effects on Blood Pressure in Diabetic Patients with Metabolic Syndrome Treated with Glimepiride

917

Hypertens Res Vol.28 (2005) No.11 p.917-924 Original Article

Giuseppe DEROSA, Arrigo Francesco Giuseppe CICERO*, Angela D’ANGELO, Antonio GADDI*, Pietro Dario RAGONESI**, Mario Nello PICCINNI, Sibilla SALVADEO, Leonardina CICCARELLI, Fabio PRICOLO, Morena GHELFI, Ilaria FERRARI, Lorenza MONTAGNA, and Roberto FOGARI

The aim of our study was to compare the long-term effect of pioglitazone and rosiglitazone on blood pressure control of diabetic patients with metabolic syndrome treated with glimepiride. We evaluated 91 type 2 diabetic patients with metabolic syndrome. All were required to have been diagnosed as diabetic for at least 6 months, and to have failed to achieve glycemic control by dietary changes and the maximum tolerated dose of the oral hypoglycemic agents sulfonylureas or metformin. All patients took a fixed dose of 4 mg/ day glimepiride. We administered pioglitazone (15 mg/day) or rosiglitazone (4 mg/day) for 12 months in a randomized, double-blind fashion, and evaluated body mass index (BMI), glycemic control, blood pressure and heart rate (HR) throughout the treatment period. A total of 87 patients completed the study and were randomized to receive double-blind treatment with pioglitazone or rosiglitazone. An increase in BMI was observed after 12 months (p<0.05) in both groups. After 9 and 12 months, there were significant decreases

in glycated hemoglobin (HbA1c), mean fasting plasma glucose (FPG), postprandial plasma glucose (PPG), fasting plasma insulin (FPI), and postprandial plasma insulin (PPI) in both treatment groups (p<0.05 at 9 months and p<0.01 at 12 months for all parameters). Furthermore, homeostasis model assessment index (HOMA index) improvement was obtained at 9 and 12 months (p<0.05 and p<0.01, respectively) in both groups. Significant systolic blood pressure (SBP) and diastolic blood pressure (DBP) improvement (p<0.05, respectively) was observed in both groups after 12 months. There were no significant changes in transaminases at any point during the study. We can conclude that the association of a thiazolinedione to the glimepiride treatment of type 2 diabetic subjects with metabolic syndrome is associated to a significant improvement in the long-term blood pressure control, related to a reduction in insulin-resistance. (Hyper- tens Res 2005; 28: 917–924)

Key Words: thiazolidinediones, pioglitazone, rosiglitazone, glimepiride, metabolic syndrome

although oral monotherapy is often initially successful, it is Introduction associated with a high secondary failure rate, which contrib- utes to the development of long-term diabetes complications Type 2 diabetes mellitus is a progressive disorder, and resulting from persistent hyperglycemia (1). For patients not

From the Internal Medicine and Therapeutics Department, University of Pavia, Pavia, Italy; *G. Descovich Atherosclerosis Study Center, Clinical Med- icine and Applied Biotechnology Department, University of Bologna, Bologna, Italy; and **Diabetes Care Unit, S. Carlo Hospital, Milano, Italy. Address for Reprints: Giuseppe Derosa, M.D., Ph.D., Department of Internal Medicine and Therapeutics, University of Pavia, P. le C. Golgi, 2−27100 Pavia, Italy. E-mail: [email protected] Received April 1, 2005; Accepted in revised form September 25, 2005. 918 Hypertens Res Vol. 28, No. 11 (2005) taking insulin, accumulating evidence suggests that combina- the effects of pioglitazone and rosiglitazone on a large set of tion therapy using oral antidiabetic agents with different biochemical parameters related to the metabolic control and mechanisms of action may be highly effective in achieving the hemorheology of these patients, as already reported in and maintaining target blood glucose levels (2). In the course Derosa et al. (11). The aim of this report was to show data rel- of the disease, the use of combinations of oral agents may ative to a secondary endpoint of our study, i.e., the effects of delay the need for insulin while maintaining glycemic con- different thiazolidinediones on blood pressure control in these trol, thus making aggressive oral treatment more acceptable patients. for many patients (3). Subjects began a controlled-energy diet (near 600 kcal Glimepiride, the latest second-generation sulfonylurea for daily deficit) based on American Diabetes Association treatment of type 2 diabetes mellitus, is a direct insulin secre- (ADA) recommendations (12) and containing 50% as carbo- tagogue; indirectly, it also increases insulin secretion in hydrates, 20% as proteins, and 30% as lipids (6% saturated), response to fuels such as glucose. Its convenient once daily with a maximum cholesterol content of 300 mg/day, and 35 g dosing may enhance compliance for diabetic patients. Glime- fiber. Each center’s standard diet advice was given by a dieti- piride is approved for monotherapy, as well as for combina- cian and/or specialist doctor. Dietitians and/or specialist doc- tion therapy with metformin and with insulin, while its tors periodically provided instruction on dietary intake possible use in association with other antihyperglycemic drug recording procedures as part of a behavior modification pro- classes is still under investigation (4). Thiazolidinediones are gram and then later used the subject’s food diaries for coun- a new class of oral hypoglycemic agents. The hypoglycemic seling. During the study, there was one behavior modification effect of the thiazolidinediones is related to their ability to session on weight-loss strategies (at baseline), one at 6 increase insulin sensitivity and, consequently, increase months, and four seminars with all patients at 3, 6, 9, and 12 peripheral glucose utilization. Although the exact mechanism months. Individuals were also encouraged to increase their of action is not completely understood, the most widely physical activity by walking briskly for 20 to 30 min, 3 to 5 accepted hypothesis is that their effect on insulin sensitivity is times per week, or by using a stationary bicycle. The recom- related to their well-known ability to bind and activate the mended changes in physical activity throughout the study nuclear peroxisomal proliferator-activated receptor-γ (PPAR- were not assessed. All patients took a fixed dose of 4 mg/day γ) (5). glimepiride (in two doses per day). In some previous studies, it was observed that the associa- Subjects were randomized for 1 year of study to also tion of sulphonylureas (rarely including glimepiride) to receive pioglitazone 15 mg/day (in a single dose) or rosiglita- rosiglitazone improve the glycemic control in type 2 diabetic zone 4 mg/day (in a single dose). They took three capsules, a patients (6−8), while the effect of pioglitazone has been tested glimepiride capsule in the morning before breakfast and more often in monotherapy (9). before dinner, and a pioglitazone or rosiglitazone capsule dur- While the metabolic effect of thiazolidinediones has ing lunch. Randomization was done using a drawing of enve- already been adequately studied, there are only limited data in lopes containing randomization codes prepared by a regard to their differential effects on blood pressure. statistician. A copy of the code was provided only to the Thus, the aim of our study was to compare the long-term responsible person performing the statistical analysis. The effect of pioglitazone and rosiglitazone on blood pressure code was only broken after database lock, but could have control of diabetic patients treated with glimepiride. been broken for individual subjects in the case of an emer- gency. Medication compliance was assessed by counting the Methods number of pills returned at the time of specified clinic visits. Pioglitazone and rosiglitazone were supplied as matching opaque white capsules in coded bottles to ensure the double- Study Design blind status of the study. At baseline, we weighed participants This multicenter, double-blind randomized trial was con- and gave them a bottle containing a supply of study medica- ducted at the Department of Internal Medicine and Therapeu- tion for at least 100 days. Throughout the study, we instructed tics of the University of Pavia, in the G. Descovich patients to take their first dose of new medication on the day Atherosclerosis Study Center of the Clinical Medicine and after they were given the study medication. A bottle contain- Applied Biotechnology Department of the University of ing the new study medication for the next treatment period Bologna, and in the Diabetes Care Unit of the S. Carlo Hospi- was given to participants on the first day of the new quarter. tal of Milano. For this study we enrolled Caucasic patients At the same time, all unused medication was retrieved for affected by type 2 diabetes and arterial hypertension, all com- inventory. All medications were provided free of charge. plying with the diagnostic criteria of metabolic syndrome as The study protocol was approved at each site by institu- defined by the National Cholesterol Education Program tional review boards and was conducted in accordance with (NCEP) (ATP III) (10). the Declaration of Helsinki. All patients provided written The primary objective of the study was to directly compare informed consent. Derosa et al: Thiazolidinedione Effects on BP in Diabetic Patients 919

Table 1. Data at Baseline in Pioglitazone and Rosiglitazone [22.8%]: losartan [3 subjects], candesartan [3 subjects], and

Group valsartan [2 subjects]; and 4 subjects, α1-antagonists [11.4%]: Glimepiride + Glimepiride + doxazosin [4 subjects]). pioglitazone rosiglitazone During the study, 45 subjects (51.7%) required antihyper- baseline baseline tensive treatment (18 subjects, ACE-inhibitors [40.0%]: ramipril [9 subjects], lisinopril [5 subjects], enalapril [3 subjects], N 45 42 and perindopril [1 subject]; 12 subjects, calcium-antagonists Sex (M/F) 21/24 22/20 [26.6%]: amlodipine [6 subjects], and lercanidipine [6 sub- ± ± Age (years) 53 6545 jects]; 10 subjects, AT II antagonists [22.2%]: losartan [4 sub- ± ± Duration of diabetes (years) 5 263 jects], and candesartan [6 subjects]; and 5 subjects, α - ± ± 1 Height (m) 1.68 0.04 1.67 0.05 antagonists [11.1%]: doxazosin [5 subjects]). ± ± Weight (kg) 68.9 3.5 67.8 3.1 Ten patients (11.4%) required addition of hydrochlorothia- 2 ± ± BMI (kg/m ) 24.4 0.8 24.3 0.7 zide; 7 patients did not require antihypertensive drugs, Data are means±SD; all group differences are nonsigniﬁcant. M, because blood pressure control was sufficient with diet regi- male; F, female; BMI, body mass index. men only. All diabetic patients were allocated to receive a fixed dose of glimepiride and a randomized dose of pioglitazone or rosiglitazone because they had poor glycemic control; in par- Patients ticular, 46 subjects (52.8%) had an insufficient glycemic con-

Diabetic patients of either sex were eligible for inclusion in trol (HbA1c >7.5%) with metformin, 27 subjects (31.0%) had the study if they had type 2 diabetes mellitus according to the an insufficient glycemic control with sulfonylureas, and ADA criteria (13). All were required to have been diagnosed exactly 13 subjects (14.9%) with gliclazide and 14 subjects as diabetic for at least 6 months, and to have failed to achieve (16.0%) with glyburide; 11 subjects (12.6%) resulted intoler- glycemic control by dietary changes and the maximum toler- ant to the metformin; and 3 subjects (3.4%) resulted intolerant ated dose of the oral hypoglycemic agents sulfonylureas or to the sulfonylureas and exactly 1 subject (1.1%) to gliclazide metformin. No patients were taking glimepiride. All patients and 2 subjects (2.2%) to glyburide. No patients were taking had a fasting C-peptide level >1.0 ng/ml. Furthermore, they lipid-lowering or antiaggregation drugs. were all hypertensive according to the World Health Organi- zation (WHO) criteria (systolic blood pressure [SBP]≥130 Assessments mmHg and diastolic blood pressure [DBP]≥85 mmHg) (14) and were all affected by metabolic syndrome (10). Suitable Before starting the study, all patients underwent an initial subjects, identified from review of case notes and/or comput- screening assessment that included a medical history, physi- erized clinic registers, were contacted personally or by tele- cal examination, vital signs, a 12-lead electrocardiogram, phone. measurements of height, weight, body mass index (BMI),

Patients with a history of ketoacidosis or with unstable or glycated hemoglobin (HbA1c), fasting plasma glucose (FPG), rapidly progressive diabetic retinopathy, nephropathy, or neu- postprandial plasma glucose (PPG), fasting plasma insulin ropathy were excluded, as were patients with impaired liver (FPI), postprandial plasma insulin (PPI), blood pressure, and function, impaired kidney function, or anemia. Patients with heart rate (HR). Changes in BMI, HbA1c, and blood pressure unstable cardiovascular conditions (e.g., New York Heart were the primary efficacy variables. Fasting plasma glucose, Association class III or IV congestive heart failure or a history PPG, and homeostasis model assessment index (HOMA of myocardial infarction or stoke) or cerebrovascular condi- index) were also used to assess efficacy. Other parameters tions within 6 months of study enrolment were excluded. have been evaluated in regard to metabolic pattern and Women who were pregnant or lactating, or who were of hemorheology, and they have been already reported else- child-bearing potential but not taking adequate contraceptive where (11). precautions were also excluded. All plasmatic parameters were determined after a 12-h Participants comprised 43 men (49.4%) and 44 women overnight fast, except PPG and PPI, which were determined 2 (50.5%) aged 47−59 years. There were no significant differ- h after lunch. Venous blood samples were taken for all ences between centers in sex distribution, age, diabetes dura- patients between 8:00 and 9:00 AM. We used plasma tion, or diabetes treatment. obtained by addition of 1 mg/ml Na2-EDTA, and centrifuged At entry, 35 subjects (40.2%) were taking antihypertensive at 3,000 × g for 15 min at 4°C. Immediately after centrifuga- drugs (13 subjects, angiotensin-converting enzyme [ACE]- tion, the plasma samples were frozen and stored at -80°C for inhibitors [37.1%]: ramipril [6 subjects], lisinopril [4 sub- no more than 3 months. All measurements were performed in jects], and enalapril [3 subjects]; 10 subjects, calcium-antag- a central laboratory. onists [28.5%]: amlodipine [7 subjects], and lercanidipine [3 BMI was calculated as weight in kg divided by the square subjects]; 8 subjects, angiotensin II [AT II] antagonists of the height in m. The estimate of insulin resistance was cal- 920 Hypertens Res Vol. 28, No. 11 (2005)

Table 2. Parameter Changes at 3, 6, 9, and 12 Months of the Study in Pioglitazone Group Glimepiride + pioglitazone Baseline 3 months 6 months 9 months 12 months BMI (kg/m2) 24.4±0.8 24.8±0.8 25.0±0.9 25.1±0.7 25.6±0.9*

HbA1c (%) 8.2±0.7 7.9±0.8 7.6±0.7 7.2±0.6* 6.8±0.8** FPG (mg/dl) 163±31 152±29 144±27 140±36* 132±25** PPG (mg/dl) 195±25 188±24 179±23 170±23* 160±22** FPI (μU/ml) 24.7±3.2 21.8±3.0 18.6±2.8 16.2±2.7* 14.8±2.6** PPI (μU/ml) 59.6±7.4 54.5±7.2 52.3±6.8 48.3±6.5* 46.1±6.3** HOMA index 10.0±3.4 8.6±3.2 7.4±3.0 6.5±2.7* 4.9±1.9** Tg (mg/dl) 156±35 150±33 144±27 130±29 121± 28*,§ SBP (mmHg) 138.2±4.4 136.5±4.6 136.1±5.0 135.3±4.5 134.4±4.2* DBP (mmHg) 87.4±4.2 86.8±4.0 85.9±3.7 85.4±3.6 84.2±3.4* HR (beats/min) 72±774±873±775±874±6 Data are means±SD; *p<0.05 vs. baseline; **p<0.01 vs. baseline; §p<0.05 vs. glimepiride + rosiglitazone. BMI, body mass index;

HbA1c, glycated hemoglobin; FPG, fasting plasma glucose; PPG, postprandial plasma glucose; FPI, fasting plasma insulin; PPI, postprandial plasma insulin; Tg, triglycerides; HOMA index, homeostasis model assessment index; SBP, systolic blood pressure; DBP, diastolic blood pressure; HR, heart rate.

Table 3. Parameter Changes at 3, 6, 9, and 12 Months of the Study in Rosiglitazone Group Glimepiride + rosiglitazone Baseline 3 months 6 months 9 months 12 months BMI (kg/m2) 24.3±0.7 24.9±0.6 25.2±0.8 25.5±0.7 25.8±0.9*

HbA1c (%) 8.0±0.8 7.8±0.9 7.3±0.8 7.0±0.6* 6.7±0.9** FPG (mg/dl) 158±33 150±30 142±29 136±29* 127±28** PPG (mg/dl) 193±28 186±26 180±25 171±24* 164±25** FPI (μU/ml) 24.1±2.8 22.4±2.7 17.9±2.6 15.1±2.5* 13.3±2.1** PPI (μU/ml) 57.4±6.8 53.6±6.9 50.5±6.6 49.1±6.5* 44.7±6.5** HOMA index 9.5±3.1 8.5±3.0 7.6±2.6 5.3±2.2* 4.2±1.7** Tg (mg/dl) 162±29 176±30 183±31 185±30 191± 32* SBP (mmHg) 137.6±3.4 136.2±3.3 135.6±3.5 135.1±4.1 133.2±3.1* DBP (mmHg) 88.8±4.6 87.6±4.5 86.9±4.2 85.4±4.0 84.0±3.8* HR (beats/min) 70±674±673±875±772±8

Data are means±SD; *p<0.05 vs. baseline; **p<0.01 vs. baseline. BMI, body mass index; HbA1c, glycated hemoglobin; FPG, fasting plasma glucose; PPG, postprandial plasma glucose; FPI, fasting plasma insulin; PPI, postprandial plasma insulin; Tg, triglycerides; HOMA index, homeostasis model assessment index; SBP, systolic blood pressure; DBP, diastolic blood pressure; HR, heart rate. culated by HOMA index using the formula FPI (μU/ml) × seated position, by using a standard mercury sphygmoma- FPG (mmol/l)/22.5, as described by Matthews and coworkers nometer (Korotkoff I and V) (Erkameter 3000, ERKA, Bad (15). Tolz, Germany) with a cuff of appropriate size. Measure-

The HbA1c level was measured by an HPLC method ments were always taken by the same investigator in the (DIAMAT, Bio-Rad, USA; normal values 4.2−6.2%), with morning before daily drug intake (i.e. ~24 h after dosing) and intra- and interassay coefficients of variation (CsV) of <2% after the subject had rested 10 min in a quiet room. Three suc- (16). Plasma glucose was assayed by the glucose-oxidase cessive blood pressure readings were obtained at 1 min inter- method (GOD/PAP, Roche Diagnostics, Mannheim, Ger- vals and averaged. BMI, HbA1c, FPG, PPG, FPI, PPI, HOMA many) with intra- and interassay CsV of <2% (17). Plasma index, SBP, DBP, and HR were evaluated at baseline and insulin was assayed with Phadiaseph Insulin RIA (Pharmacia, after 3, 6, 9, and 12 months. Uppsala, Sweden) by using a second antibody to separate the In order to evaluate the tolerability assessments, all adverse free and antibody-bound 125I-insulin (intra- and interassay events were recorded. Blood pressure values were measured CsV: 4.6 and 7.3%, respectively) (18). Blood pressure mea- by physicians not belonging to the study to preserve study surements were obtained from each patient (right arm) in the blindness by experimenters. Derosa et al: Thiazolidinedione Effects on BP in Diabetic Patients 921

0 7

-1

6 -2

Pioglitazone -3 Rosiglitazone 5 -4 * * Blood Pressure (mmHg) -5 4 * DBP (mmHg) -6 * SBP DBP 3 rosiglitazone Fig. 1. Change in SBP and DBP from baseline to 12 months in patients receiving pioglitazone or rosiglitazone. Values 2 pioglitazone are the mean±SD. *p<0.05, change from baseline. SBP, 7 8 9 10 11 12 13 systolic blood pressure; DBP, diastolic blood pressure. FPI (microU/ml)

Statistical Analysis Fig. 2. Correlation between FPI decrease and DBP decrease in pioglitazone and rosiglitazone group. Pioglita- An intention-to-treat analysis was conducted in patients who zone: DBP delta =0.124x + 0.299; p=0.005. Rosiglitazone: had received at least one dose of study medication and had a DBP delta =-0.789x - 0.266; p=0.013. DBP, diastolic subsequent efficacy observation. Patients were included in blood pressure; FPI, fasting plasma insulin. the safety analysis if they had received one dose of trial medication after randomization and had a subsequent safety observation. The null hypothesis that the expected mean SBP the patient population at study entry, shown in Table 1, were and DBP change from baseline to the end of 12 months of similar in the two treatment groups. double-blind treatment did not differ significantly between pioglitazone and rosiglitazone treatments was tested using Body Mass Index Value analysis of variance (ANOVA) and analysis of covariance (ANCOVA) models (19). Similar analyses were applied to No BMI change was observed after 3, 6, or 9 months in both the other parameters. The statistical significance of the inde- groups, while a significant BMI increase was present after 12 pendent effects of treatments on the other parameters was months compared to the baseline (p<0.05) in both groups. determined by ANCOVA. A one-sample t test was used to There was no difference in BMI value between the pioglita- compare values obtained before and after treatments adminis- zone and rosiglitazone group. The results are reported in tration; and two-sample t tests were used for between-group detail in Tables 2 and 3. comparison. Statistical analysis of data was performed by means of the SPSS statistical software package for Windows Glycemic Control (version 11.0; Chicago, USA); data are presented as the mean±SD. For all statistical analysis, p<0.05 was considered No HbA1c change was observed after 3 and 6 months in both statistically significant. groups, while a significant HbA1c decrease was obtained after 9 (p<0.05) and 12 (p<0.01) months in both groups. No sig- Results nificant FPG and PPG variation was present at 3 and 6 month in both groups. After 9 and 12 months, the mean FPG and PPG levels were significantly decreased in both groups Study Sample (p<0.05 and p<0.01, respectively) (Tables 2 and 3). Fasting A total of 91 patients were enrolled in the trial. Of these, 87 plasma insulin and PPI did not show any significant change completed the study: 45 of the 87 (51.7%) were randomized after 3 or 6 months, while a significant decrease was observed to receive pioglitazone and 42 (48.2%) to receive rosiglita- at 9 and 12 months (p<0.05 and p<0.01, respectively) com- zone in a double-blind fashion. There were 4 patients (2 males pared to the baseline value in both groups. Furthermore, and 2 females) who did not complete the study and the rea- HOMA index improvement was obtained only at 9 and 12 sons for premature withdrawal included protocol violation, months (p<0.05 and p<0.01, respectively) compared to the loss to follow-up, and non-compliance. The characteristics of baseline value in both groups. 922 Hypertens Res Vol. 28, No. 11 (2005)

ficiently low to be considered safe for the diabetic patient: in Blood Pressure Effects fact, it is important to remember that the third report of the No SBP or DBP change was observed in either group after 3, NCEP Expert Panel on the Detection, Evaluation and Treat- 6, or 9 months. Significant SBP and DBP improvement ment of High Blood Cholesterol in Adults has defined the dia- (p<0.05, respectively) was present in both groups after 12 betic disease as a coronary heart disease risk equivalent (10). months compared to the baseline values (Figs. 1 and 2). No Because insulin-resistance syndrome is also considered a pri- significant HR variation was obtained during the study in mary cause of hypertension (23−25) and cardiovascular dis- either group (Tables 2 and 3). ease (26), a specific treatment to improve this syndrome is also required. Indeed, hypertension control represents the most important Safety intervention, since it limits cardiovascular events far more Of the 87 patients who completed the study, 6.6% (3/45) of effectively than tight glycemic control (20), but not all antihy- patients in the pioglitazone group and 11.9% (5/42) of pertensive treatments have the same impact on the metabolic patients in the rosiglitazone group had side effects (not signif- control of diabetic patients (27). To the best of our knowl- icant). In the pioglitazone group, 1 patient had aspartate ami- edge, this is the first study to compare the long-term effects of notransferase (AST) and alanine aminotransferase (ALT) glycemic control by thiazolidinediones on the blood pressure values that increased to 1.5 times the upper limit of normal control of type 2 diabetic subjects. (<40 U/l), but these values regressed after 15 days and It has already been established that thiazolidinediones have returned to within the limit of normal; 1 patient reported tran- an optimal effect on glycemic control in diabetic subjects sient headache (only for 5 days from the start of therapy); and (28), and this was further confirmed in the present study: both 1 patient reported transient flatulence (only for 3 days from combination treatment with glimepiride and pioglitazone and the start of therapy). In the rosiglitazone group, 2 patients had that with glimepiride and rosiglitazone significantly AST and ALT values that increased to 2.0 times the upper improved the glycemic control in the studied subjects. We limit of normal (<40 U/l), but these values regressed after 15 observed a 1.3% improvement in HbA1c plasma level days and returned to within the limit of normal; 2 patients (p<0.01), a 19.3% in FPG (p<0.01), a 16.3% in PPG reported transient flatulence (only for 2 days and only for 1 (p<0.01), a 42.4% in FPI (p<0.05) and a 23.3% in PPI day from the start of therapy, respectively); and 1 patient (p<0.05), without significant differences between treatment reported transient headache (only for 2 days from the start of groups. Furthermore, we found a slight, but statistically sig- therapy). nificant reduction in both SBP and DBP in both the pioglita- Altogether, we did not have statistically significant changes zone- and rosiglitazone-treated groups, while no significant in transaminases. In particular, in pioglitazone group, AST change in HR was observed. In contrast to what has already and ALT values were 24±8 and 23±9 U/l at baseline, respec- been observed for some metabolic parameters (more specifi- tively, and 26±9 and 24±10 U/l at 12 months, respectively. cally, the different effect of the two available thiazolidinedi- In rosiglitazone group, AST and ALT values were 25±9 and ones on plasma lipid pattern), a nonsignificant difference has 22±8 U/l at baseline, respectively, and 27±10 and 24±9 U/l been noted in regard to blood pressure improvement under at 12 months, respectively. thiazolidinediones treatment (11). Of course, we recognize that in this case the lack of a control group renders our results Discussion slightly weaker. The observed result was confirmed even repeating statistical analysis by classes of antihypertensive Coronary artery disease (CAD) causes much of the serious drug assumed by the patients. These data could suggest an morbidity and mortality in patients with diabetes who have a independent effect of concomitant blood pressure lowering 2- to 4-fold in the risk of CAD (20). Hypertension, hypercho- drug eventually co-assumed. A similar antihypertensive lesterolemia and low high-density lipoprotein cholesterol effect has been seen even with other drugs affecting the lipid (HDL-C) values exponentially increase the cardiovascular metabolism, such as statins and fibrates (29, 30). The antihy- risk profile of type 2 diabetic patients, among whom these pertensive effect of thiazolidinediones appears to be mainly risk factors are highly prevalent. On the other hand, an inten- related to the decrease in insulin-resistance and the conse- sive treatment of all modifiable risk factors significantly quent improvement of endothelial function (31, 32). In our improves the prognosis of these patients (1). In particular, an study, we observed a correlation between FPI decrease and adequate antihypertensive treatment has clearly been show to DBP decrease in both groups (Fig. 2). The observed antihy- prevent or to slow the diabetes-related renal damage (21). pertensive effect of thiazolidinediones was probably attenu- However, beyond that evidence, the control rate of blood ated by the concomitant treatment with glimepiride, which is pressure in these patients remains very low (22). already known to have a long-term improving effect on blood Therefore, it is not sufficient to simply treat each risk fac- pressure control (33). tor, but rather the treatment must be very intensive in order to On the basis of our results, we can conclude that the associ- obtain both blood pressure and cholesterolemia at levels suf- ation of a thiazolidinedione to the glymepiride treatment of Derosa et al: Thiazolidinedione Effects on BP in Diabetic Patients 923 type 2 diabetic subjects is associated to a significant improve- 151−183. ment in the long-term blood pressure control related to a 15. Matthews DR, Hosker JP, Rudenski AS, Naylor BA, reduction in insulin-resistance. The choice of the best mole- Treacher DF, Turner RC: Homeostasis model assessment: β cule to prescribe has to be driven by the drug effects on other insulin resistance and -cell function from fasting plasma parameters such as body weight, plasma lipid levels and pro- glucose and insulin concentrations in man. Diabetologia 1985; 28: 412−419. thrombotic risk factors. 16. Bunn HF, Gabbay HK, Gallop PM: The glycosylation of haemoglobin. Relevance to diabetes mellitus. Science 1978; References 200: 21−27. 17. European Diabetes Policy Group: A desktop guide to type 2 1. Expert Committee on the Diagnosis and Classification of diabetes mellitus. Diabet Med 1999; 16: 716−730. Diabetes Mellitus: American Diabetes Association: clinical 18. Heding LG: Determination of total serum insulin (IRI) in practice recommendations 2002. Diabetes Care 2002; 25 insulin-treated diabetic patients. Diabetologia 1972; 8: (Suppl): S1−S147. 260−266. 2. Charpentier G, Fleury F, Kabir M, Vaur L, Halimi S: 19. Winer BJ: Statistical Principles in Experimental Design, Improved glycaemic control by addition of glimepiride to 2nd ed. New York, McGraw-Hill, 1971. metformin monotherapy in type 2 diabetic patients. Diabet 20. Beckman JA, Creager MA, Libby P: Diabetes and athero- Med 2001; 18: 828−834. sclerosis. Epidemiology, pathophysiology, and manage- 3. Riddle M: Combining sulfonylureas and other oral agents. ment. JAMA 2002; 287: 2570−2581. Am J Med 2000; 108 (Suppl 6a): 15S−22S. 21. Vivian EM, Rubinstein GB: Pharmacologic management of 4. McCall AL: Clinical review of glimepiride. Expert Opin diabetic nephropathy. Clin Ther 2002; 24: 1741−1756. Pharmacother 2001; 2: 699−713. 22. Katayama S, Inaba M, Morita T, Awata T, Shimamoto K, 5. Ovalle F, Ovalle-Berúmen JF: Thiazolidinediones: a review Kikkawa R: Blood pressure control in Japanese hyperten- of their benefits and risks. South Med J 2002; 95: 1187− sives with or without type 2 diabetes mellitus. Hypertens 1194. Res 2000; 23: 601−605. 6. Wolffenbuttel BH, Gomis R, Squatrito S, Jones NP, Pat- 23. Nishina M, Kikuchi T, Yamazaki H, Kameda K, Hiura M, wardhan RN: Addition of low-dose rosiglitazone to sulpho- Uchiyama M: Relationship among systolic blood pressure, nylurea therapy improves glycaemic control in type 2 serum insulin and leptin, and visceral fat accumulation in diabetic patients. Diabet Med 2000; 17: 40−47. obese children. Hypertens Res 2003; 26: 281−288. 7. Vongthavaravat V, Wajchenberg BL, Waitman JN, et al, 24. Anan F, Takahashi N, Ooie T, Yufu K, Saikawa T, Yoshi- 125 Study Group: An international study of the effects of matsu H: Role of insulin resistance in nondipper essential rosiglitazone plus sulphonylurea in patients with type 2 dia- hypertensive patients. Hypertens Res 2003; 26: 669−676. betes. Curr Med Res Opin 2002; 18: 456−461. 25. Hirose H, Saito I, Kawabe H, Saruta T: Insulin resistance 8. St John Sutton M, Rendell M, Dandona P, et al: A compari- and hypertension: seven-year follow-up study in middle- son of the effects of rosiglitazone and glyburide on cardio- aged Japanese men (the KEIO study). Hypertens Res 2003; vascular function and glycemic control in patients with type 26: 795−800. 2 diabetes. Diabetes Care 2002; 25: 2058−2064. 26. Harano Y, Suzuki M, Shinozaki K, et al: Clinical impact of 9. Pavo I, Jermendy G, Varkonyi TT, et al: Effect of pioglita- insulin resistance syndrome in cardiovascular diseases and zone compared with metformin on glycemic control and its therapeutic approach. Hypertens Res 1996; 9 (Suppl 1): indicators of insulin sensitivity in recently diagnosed S81−S85. patients with type 2 diabetes. J Clin Endocrinol Metab 27. Derosa G, Ragonesi PD, Mugellini A, Ciccarelli L, Fogari 2003; 88: 1637−1645. R: Effects of telmisartan compared with eprosartan on 10. National Cholesterol Education Program (NCEP) Expert blood pressure control, glucose metabolism and lipid profile Panel: Summary of the third report of the NCEP expert in hypertensive, type 2 diabetic patients: a randomized, dou- panel on detection, evaluation, and treatment of high blood ble-blind, placebo-controlled 12-month study. Hypertens cholesterol (Adult Treatment Panel III). JAMA 2001; 285: Res 2004; 27: 457−464. 2486−2497. 28. Boyle PJ, King AB, Olansky L, et al: Effects of pioglita- 11. Derosa G, Cicero AFG, Gaddi A, et al: Metabolic effects of zone and rosiglitazone on blood lipid levels and glycemic pioglitazone and rosiglitazone in patients with diabetes and control in patients with type 2 diabetes mellitus: a retrospec- metabolic syndrome treated with glimepiride: a twelve- tive review of randomly selected medical records. Clin Ther month, multicenter, double-blind, randomized, controlled, 2002; 24: 378−396. parallel-group trial. Clin Ther 2004; 26: 744−754. 29. Kim JI, Tsujino T, Fujioka Y, Saito K, Yokoyama M: 12. American Diabetes Association: Nutrition recommenda- Bezafibrate improves hypertension and insulin sensitivity in tions and principles for people with diabetes mellitus (posi- humans. Hypertens Res 2003; 26: 307−313. tion statement). Diabetes Care 2001; 24 (1): S44−S47. 30. Kanbay M, Yildirir A, Bozbas H, et al: Statin therapy helps 13. American Diabetes Association: Screening for diabetes to control blood pressure levels in hypertensive dyslipi- (position statement). Diabetes Care 2001; 24 (1): S21−S24. demic patients. Ren Fail 2005; 27: 297−303. 14. Guidelines Subcommittee: 1999 World Health Organiza- 31. Shargorodsky M, Wainstein G, Gavish E, Leibovitz Z, tion−International Society of Hypertension Guidelines for Matas D, Zimlichman R: Treatment with rosiglitazone the management of hypertension. J Hypertens 1999; 17: reduces hyperinsulinemia and improves arterial elasticity in 924 Hypertens Res Vol. 28, No. 11 (2005)

patients with type 2 diabetes mellitus. Am J Hypertens 33. Derosa G, Mugellini A, Ciccarelli L, Crescenzi G, Fogari R: 2003; 16: 617−622. Comparison between repaglinide and glimepiride in 32. Fullert S, Schneider F, Haak E, et al: Effects of pioglitazone patients with type 2 diabetes mellitus: a one-year, random- in nondiabetic patients with arterial hypertension: a double- ized, double-blind assessment of metabolic parameters and blind, placebo-controlled study. J Clin Endocrinol Metab cardiovascular risk factors. Clin Ther 2003; 25: 472−484. 2002; 87: 5503−5506.